Syléa

• Outcome: Review
• Latest audit: Sep 16, 2026 · 6:54 AM UTC
• Version: 0.1.0

Overview

Syléa is a coherent local coaching skill, but it persistently saves sensitive personal and well-being information in plaintext local files with broad triggers and limited privacy controls.

Install only if you are comfortable with a local long-term coaching journal under ~/.sylea/ that may include identity details, goals, dilemmas, routines, and well-being scores. Before using it, consider restricting directory and file permissions, avoiding highly sensitive disclosures, and manually reviewing or deleting stored Markdown files when you no longer need them.

A.I.G

Vulnerability Patterns

• Insecure Skill Coding Practices: Finds exploitable flaws such as hardcoded secrets or command injection
• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally

Findings (1)

T09 · Insecure Skill Coding Practices

Warning

• Location: SKILL.md:44

Finding

Plaintext Storage of Sensitive Personal Data Without Explicit Access Controls

Content

View full analysis

Créé le ## Identité - Âge: - Profession: <...> - Ville: <...> - Situation: <...> ## Objectif principal - Catégorie: - Deadline: - Heures/jour engagées: ## Forces - - - ## Baseline bien-être () | Énergie | Motivation | Stress | Social | |---------|------------|--------|--------| | X/10 | X/10 | X/10 | X/10 | ## Budget temps (h/jour) - Sommeil: X - Travail: X - Loisirs: X - Transport: X - Objectif: X ``` ``` ### Technical Analysis The skill directs the agent to collect and persistently store personally s ...[truncated 2140 chars]

Remediation

View remediation

SkillSpector

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Rogue Agent: Self-Modification, Session Persistence
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration

Findings (7)

Natural-Language Policy Violations

Medium

• Category: Not specified by scanner
• Confidence: 95% confidence

Finding

The changelog states 'French-first responses' with fallback to English and other languages, which implies a default language preference that may be imposed on users. Under the policy rule, forcing or defaulting to a specific language without explicit opt-in can be a natural-language policy violation unless the locale constraint is clearly justified.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium

• Category: Not specified by scanner
• Confidence: 95% confidence

Finding

Line L45 explicitly markets the skill as "Français natif" and the entire user-facing documentation is framed only in French, with no indication that users may opt into another language. Under the policy, forcing a specific language without user choice is a natural-language policy concern unless the locale restriction is clearly justified.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium

• Category: Not specified by scanner
• Confidence: 95% confidence

Finding

The skill states that data lives locally, but the user-facing behavior does not prominently warn that highly sensitive personal information will be stored persistently in plaintext files, including identity, psychological baseline, goals, dilemmas, and daily well-being logs. In a coaching/well-being context, this omission materially increases privacy risk because users may disclose sensitive data assuming the exchange is ephemeral.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium

• Category: Not specified by scanner
• Confidence: 95% confidence

Finding

The activation rules are broad and match common everyday phrases about stress, choices, goals, and check-ins, which can cause the skill to trigger in situations where the user did not explicitly ask for persistent coaching behavior. In this skill, accidental activation is more sensitive because it can lead to collection and storage of intimate personal, psychological, and routine data under ~/.sylea/ without a strong per-use consent gate.

Content

No source excerpt is available for this finding.

Session Persistence

Medium

• Category: Rogue Agent
• Confidence: 97% confidence

Finding

The skill is explicitly designed for session persistence by creating ~/.sylea/ and maintaining long-lived records across profile, goals, dilemmas, and check-ins. Persistent storage is especially risky here because the data includes sensitive mental-health-adjacent and life-planning information that could be exposed to other local users, backups, logs, or compromise of the host environment.

Content

Scanner excerpt · SKILL.md (reported line 39)May include surrounding context.

md

- Asks for a **daily bilan** — keywords: *"how was my day"*, *"bilan journée"*, *"check-in"*
- Wants **progress tracking** on an existing objective

Respond in the user's language (French if they write French, English otherwise). Default to French when ambiguous.

## Setup — first use

Session Persistence

Medium

• Category: Rogue Agent
• Confidence: 96% confidence

Finding

Saving the full dilemma analysis, including the user's reasoning and recommendation, creates a durable record of potentially intimate personal conflicts, relationships, finances, career decisions, and emotional state. Because this content can be highly sensitive and inferential, its retention increases privacy and safety risks well beyond ordinary note-taking, especially when stored in predictable plaintext paths.

Content

Scanner excerpt · SKILL.md (reported line 138)May include surrounding context.

md

- If `well-being < 5` on the winner → add *"⚠️ attention au burn-out sur cette option"*

### Step 6. Save the analysis
Write to `~/.sylea/dilemmas/YYYY-MM-DD-<slug>.md` with the full table, the user's reasoning if shared, and the recommendation.

Never decide FOR the user. Surface the analysis; they pick.

Scope Creep

Low

• Category: Excessive Agency
• Confidence: 70% confidence

Finding

Skill's behavior or capabilities extend beyond its stated purpose. Scope creep allows an agent to perform actions unrelated to its documented functionality, increasing the attack surface.

Content

Scanner excerpt · LICENSE.md (reported line 12)May include surrounding context.

md

the Software, and to permit persons to whom the Software is furnished to do so.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS
FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN

Static analysis

No suspicious patterns detected.